‘What if AI became a virtual copywriter? Just imagine what would happen if your content marketer had access to your best-performing pieces and was able to produce original content that is just as good as a professional. Yes, this is one future scenario of the content marketing revolution”.
That is what AI itself has to say about the future of content marketing. The above introduction was created using Copi.ai and it does sound promising. But if you were not told that it was AI-generated text, you likely would have believed it was written by a human. There is already a lot of content on the internet, in different forms, that has been generated using artificial intelligence. This content is customized as per a brands’ needs, but may soon be personalized according to each customer’s individualized purchase behavior.
Artificial intelligence is quickly developing to become a marketer’s best friend. But what is AI exactly?
Artificial Intelligence
Artificial intelligence is a technology that is trained to study data and respond in a way humans would. Brands are increasingly adopting this technology to automate processes so that their marketing teams can devote more time to the strategic aspects of the business. There are already many Martech tools available that help brands create content in the form of blogs, press releases, ad copy, e-mail subject lines, commercials, movies, etc.
In September 2020, The Guardian published an article that was written entirely by AI using GPT-3, a machine learning language model developed by OpenAI. The model was fed a set of instructions by the publication along with the introduction to give it a head start. The article appears to be no less than a human creation. It was later disclosed that the Guardian generated eight different essays using the same introduction and guidelines, picked certain parts from each, edited it, and created this piece. Although the final copy had some human editing, it remains an exceptional piece of content and helps to illustrate the potential of AI in writing content.
If you have run advertising campaigns, then you must know that artificial intelligence has been with you throughout the process. Targeting the right audience with the right message, at the right moment is what the machine has been helping you with your campaigns. But what if we told you that AI can create ads or develop scripts for ads right from scratch? Yes, you read that right! In 2019, Lexus launched an ad that was entirely scripted by AI.
By now, you might be convinced that AI does a great job in writing short-form content. But it doesn’t stop there. In 2016, a short film called Sunspring was released. It is a 9-minute-long film that showcases three individuals living in the future with a hint of love triangle to it. It appears to be like any other sci-fi movie apart from the fact that it was scripted by AI. This script was developed using Benjamin – the name of the AI tool that was given to it by the machine itself. Benjamin was fed with a huge amount of data from sci-fi movies in order to come up with the script. While the story has gaps in sensibility, technology is evolving quickly and we can expect that in the future an entire movie might be written by AI.
In these examples, we find one thing in common which is feeding data to the AI in order to get the intended output in the form of relevant content. This is similar to what content writers do i.e., first, read a plethora of information about the topic they plan to write on; second, understand and process the info and third, create an exclusive piece of writing. The various AI tools available work in a similar fashion. The difference is that they use artificial intelligence instead of human intelligence. So, to read the data, AI software uses Natural Language Processing (NLP), Natural Language Understanding (NLU), and to finally generate content that resonates with the brand and its audience, Natural Language Generation (NLG).
Generative Pre-trained Transformer (GPT-3)
The AI that the various tools uses to generate content is called Generative Pre-trained Transformer (GPT). Currently, the most recent AI model is the third generation of GPT, now commonly known as GPT-3. It was initially developed by OpenAI, a research and development company founded by Elon Musk that aims to create autonomous systems using artificial intelligence.
GPT-3 is trained by using over 175 billion parameters. These parameters help the model to create content and respond to descriptions and questions in a way humans would. GPT-3 has been trained to mimic being sentient, have a sense of humour, differentiate between a truth and a lie, and also provide factual as well as subjective content. It is smart enough to not just write complete ad scripts but also share some realistic investment advice.
In the article published by The Guardian mentioned earlier, we saw that although it was the machine that created the final output, humans were required in the process to edit the final draft. However, reduced human effort will be required in the upcoming fourth generation of GPT (i.e. GPT-4). GPT-4 is expected to be launched in the third quarter of 2022. It will be bigger in size than GPT-3, and may have up to 280 billion parameters. GPT-4 will enable online tools to share content that will appear to be more human-like.
What does the future hold?
In the years to come, GPT will undoubtedly advance. Marketers will start to rely on AI to not just develop content, but also provide feedback on a product or service based on the consumer data it has studied.
We predict we will also have designated digital assistants, created by using Generative Adversarial Networks (GANs), playing the role of product consultants in various organizations. These CGI digital creations will be given a face of a person who does not even exist who will read out a script written by the GPT-X generation, display human-like emotions and be a part of a scene created and directed by AI and altogether form an ad.
Further, we can expect technology-driven organizations of the future to further automate social media marketing. The AI will know which type of content is performing well and it will create more of that kind of content, targeted toward a specific audience. The only thing that social media marketers may be required to do is approve, reject, or edit the ad budget suggested by AI and extract weekly, monthly, quarterly, and yearly content and campaign reports to be presented to the management.
Engaging with existing customers and keeping them up to date may also be available at the click of a button. Imagine logging in to a portal, selecting a target group, based on the campaign you want to run, from the list of buyer personas created by the machine for you, and simply click, to send a customized newsletter to each of the members of your target audience. This newsletter will be tailored to individual consumers, based on their activity not just on your website, but also from data collected in your virtual and physical store.
However, the future of content creation also presents challenges. How will consumers determine the authenticity of the content? Does that matter? With AI writing, it will become difficult to differentiate between original and machine-generated content. Also, with so many tools available online for free, knowing the difference between fake and real communication will be a major problem. As a result, brands will have to constantly find ways to establish trust amongst their audience.
What can copy and content writers look forward to?
With all the examples that we have seen so far in this article, it can be rightly said that nowhere was AI-generated content used ‘As-is’. The Guardian article was published after editing eight different essays while the Lexus ad and the movie Sunspring required human intervention in terms of the direction of the video to make sense of the machine-generated content. Although GPT-4 will be trained by using much more data, there is still a lot of work required until perfect content generation is achieved. Until then, brands will highly depend on content and copywriters to polish GPT-3 content to make it sound in line with the brand personality as well as the tone.
Until now, when hiring content marketers and copywriters, firms looked for key skills of researching and writing in the candidate. However, working knowledge of AI will be a key skill moving forward. The basic criteria would include candidates who research well, understand AI, and possess the technical knowledge of feeding data into models and getting them to generate the right content. So, it is all the more essential that current and future content marketers brush up their technical skills to stay ahead of the competition.
Written by: